From: Peter Xu <peterx@redhat.com>
To: qemu-devel@nongnu.org
Cc: qemu-block@nongnu.org, "Fabiano Rosas" <farosas@suse.de>,
"Dr . David Alan Gilbert" <dave@treblig.org>,
"Eric Blake" <eblake@redhat.com>, "Kevin Wolf" <kwolf@redhat.com>,
"Vladimir Sementsov-Ogievskiy" <vsementsov@yandex-team.ru>,
"Daniel P . Berrangé" <berrange@redhat.com>,
"Andrey Drobyshev" <andrey.drobyshev@virtuozzo.com>,
peterx@redhat.com, "Stefan Hajnoczi" <stefanha@redhat.com>
Subject: [PATCH RFC 11/11] tests/qtest/migration: Test successive migrations
Date: Tue, 3 Dec 2024 19:51:38 -0500 [thread overview]
Message-ID: <20241204005138.702289-12-peterx@redhat.com> (raw)
In-Reply-To: <20241204005138.702289-1-peterx@redhat.com>
From: Fabiano Rosas <farosas@suse.de>
Add a framework for running migrations back and forth between two
guests (aka ping-pong migration). We have seen a couple of bugs that
only reproduce when a guest is migrated and the destination machine is
used as the source of a new migration.
Add a simple test that does 2 migrations and another test that uses
the late-block-activate capability, which is one area where a bug has
been found.
Note that this does not reuse any of the existing tests
(e.g. test_precopy_common), because there is too much ambiguity
regarding how to handle the hooks, args->result, stop/continue, etc.
Signed-off-by: Fabiano Rosas <farosas@suse.de>
Link: https://lore.kernel.org/r/20241125144612.16194-6-farosas@suse.de
Signed-off-by: Peter Xu <peterx@redhat.com>
---
tests/qtest/migration-test.c | 121 +++++++++++++++++++++++++++++++++++
1 file changed, 121 insertions(+)
diff --git a/tests/qtest/migration-test.c b/tests/qtest/migration-test.c
index 74645f749c..4955b3b9eb 100644
--- a/tests/qtest/migration-test.c
+++ b/tests/qtest/migration-test.c
@@ -1819,6 +1819,104 @@ finish:
test_migrate_end(from, to, args->result == MIG_TEST_SUCCEED);
}
+static void migrate_dst_becomes_src(QTestState **from, QTestState **to,
+ QTestMigrationState *src_state,
+ QTestMigrationState *dst_state)
+{
+ *from = *to;
+
+ *src_state = (QTestMigrationState) { };
+ src_state->serial = dst_state->serial;
+ qtest_qmp_set_event_callback(*from, migrate_watch_for_events, src_state);
+
+ src_state->stop_seen = dst_state->stop_seen;
+}
+
+static void test_precopy_ping_pong_common(MigrateCommon *args, int n,
+ bool late_block_activate)
+{
+ QTestState *from, *to;
+ bool keep_paused = false;
+
+ g_assert(!args->live);
+
+ g_test_message("Migrating back and forth %d times", n);
+
+ for (int i = 0; i < n; i++) {
+ g_test_message("%s (%d/%d)", i % 2 ? "pong" : "ping", i + 1, n);
+
+ if (test_migrate_start(&from, &to, args->listen_uri, &args->start)) {
+ return;
+ }
+
+ if (late_block_activate) {
+ migrate_set_capability(from, "late-block-activate", true);
+ migrate_set_capability(to, "late-block-activate", true);
+
+ /*
+ * The late-block-activate capability expects that the
+ * management layer will issue a qmp_continue command on
+ * the destination once the migration finishes. In order
+ * to test the capability properly, make sure the test is
+ * not issuing spurious continue commands.
+ */
+ keep_paused = true;
+ }
+
+ if (!src_state.stop_seen) {
+ wait_for_serial(src_state.serial);
+ }
+
+ /* because of !args->live */
+ qtest_qmp_assert_success(from, "{ 'execute' : 'stop'}");
+ wait_for_stop(from, &src_state);
+
+ migrate_ensure_converge(from);
+ migrate_qmp(from, to, args->connect_uri, args->connect_channels, "{}");
+
+ wait_for_migration_complete(from);
+ wait_for_migration_complete(to);
+
+ /* note check_guests_ram() requires a stopped guest */
+ check_guests_ram(to);
+
+ if (keep_paused) {
+ /*
+ * Nothing issued continue on the destination, reflect
+ * that the guest is paused in the dst_state.
+ */
+ dst_state.stop_seen = true;
+ } else {
+ qtest_qmp_assert_success(to, "{ 'execute' : 'cont'}");
+ wait_for_serial(dst_state.serial);
+ dst_state.stop_seen = false;
+ }
+
+ /*
+ * Last iteration, let the guest run to make sure there's no
+ * memory corruption. The test_migrate_end() below will check
+ * the memory one last time.
+ */
+ if (i == n - 1) {
+ qtest_qmp_assert_success(to, "{ 'execute' : 'cont'}");
+ wait_for_serial(dst_state.serial);
+ dst_state.stop_seen = false;
+ break;
+ }
+
+ /*
+ * Prepare for migrating back: clear the original source and
+ * switch source & destination.
+ */
+ migrate_cleanup(from, NULL);
+ migrate_dst_becomes_src(&from, &to, &src_state, &dst_state);
+
+ args->start.only_target = true;
+ }
+
+ test_migrate_end(from, to, true);
+}
+
static void file_dirty_offset_region(void)
{
g_autofree char *path = g_strdup_printf("%s/%s", tmpfs, FILE_TEST_FILENAME);
@@ -3772,6 +3870,24 @@ static void test_migrate_dirty_limit(void)
test_migrate_end(from, to, true);
}
+static void test_precopy_ping_pong(void)
+{
+ MigrateCommon args = {
+ .listen_uri = "tcp:127.0.0.1:0",
+ };
+
+ test_precopy_ping_pong_common(&args, 2, false);
+}
+
+static void test_precopy_ping_pong_late_block(void)
+{
+ MigrateCommon args = {
+ .listen_uri = "tcp:127.0.0.1:0",
+ };
+
+ test_precopy_ping_pong_common(&args, 2, true);
+}
+
static bool kvm_dirty_ring_supported(void)
{
#if defined(__linux__) && defined(HOST_X86_64)
@@ -4055,6 +4171,11 @@ int main(int argc, char **argv)
}
}
+ migration_test_add("/migration/precopy/ping-pong/plain",
+ test_precopy_ping_pong);
+ migration_test_add("/migration/precopy/ping-pong/late-block-activate",
+ test_precopy_ping_pong_late_block);
+
ret = g_test_run();
g_assert_cmpint(ret, ==, 0);
--
2.47.0
next prev parent reply other threads:[~2024-12-04 0:53 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-12-04 0:51 [PATCH RFC 00/11] migration/block: disk activation rewrite Peter Xu
2024-12-04 0:51 ` [PATCH RFC 01/11] migration: Add helper to get target runstate Peter Xu
2024-12-04 0:51 ` [PATCH RFC 02/11] migration/block: Make late-block-active the default Peter Xu
2024-12-04 0:51 ` [PATCH RFC 03/11] migration/block: Apply late-block-active behavior to postcopy Peter Xu
2024-12-04 0:51 ` [PATCH RFC 04/11] migration/block: Fix possible race with block_inactive Peter Xu
2024-12-04 0:51 ` [PATCH RFC 05/11] migration/block: Merge block reactivations for fail/cancel Peter Xu
2024-12-04 0:51 ` [PATCH RFC 06/11] migration/block: Extend the migration_block_* API to dest side Peter Xu
2024-12-04 14:49 ` Peter Xu
2024-12-04 0:51 ` [PATCH RFC 07/11] migration/block: Apply the migration_block_* API to postcopy Peter Xu
2024-12-04 0:51 ` [PATCH RFC 08/11] tests/qtest/migration: Move more code under only_target Peter Xu
2024-12-04 0:51 ` [PATCH RFC 09/11] tests/qtest/migration: Don't use hardcoded strings for -serial Peter Xu
2024-12-04 19:11 ` Fabiano Rosas
2024-12-04 0:51 ` [PATCH RFC 10/11] tests/qtest/migration: Support cleaning up only one side of migration Peter Xu
2024-12-04 0:51 ` Peter Xu [this message]
2024-12-04 17:25 ` [PATCH RFC 00/11] migration/block: disk activation rewrite Peter Xu
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20241204005138.702289-12-peterx@redhat.com \
--to=peterx@redhat.com \
--cc=andrey.drobyshev@virtuozzo.com \
--cc=berrange@redhat.com \
--cc=dave@treblig.org \
--cc=eblake@redhat.com \
--cc=farosas@suse.de \
--cc=kwolf@redhat.com \
--cc=qemu-block@nongnu.org \
--cc=qemu-devel@nongnu.org \
--cc=stefanha@redhat.com \
--cc=vsementsov@yandex-team.ru \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).